DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA . This action is in response to instant application filed on 06/24/2024. Claims 1-20 have been examined.
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 06/24/2024 is being considered by the examiner.
Claim Objections
Claim 1 is objected to because of the following informalities: "wherein thethe one or". Appropriate correction is required. Claim 20 is objected to because of the following informalities: claim 20 appears to be a part of claim 19 due to the lack of spacing. Claims 5 and 16 are objected to because of the following informalities: “analyzing and predicting a performance the at least one candidate data”, there should be an “of” after performance.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Regarding claim 1:
Step 1: The claim recites a method which falls into the statutory category of process.
Step 2A Prong 1: The claim recites multiple mental processes: determining whether the at least one candidate data driven model meets a predefined prediction and tuning at least one hyper-parameter and at least one data driven technique upon determining that the at least one candidate data driven model does not meet the predefined prediction. A human being can mentally determine whether a model has met some goal and then decide to change the parameters of the model all in the human mind.
Step 2A Prong 2: Claim 1 does not integrate the abstract idea into a practical application since the additional elements of: obtaining at least one candidate data driven model from a plurality of candidate data driven models and deploying the at least one candidate data driven model to at least one target deployment environment upon determining that the at least one candidate data driven model meets the predefined prediction are considered insignificant extra-solution activity: data gathering (MPEP 2106.05(g)). Meanwhile, the additional element of an electronic device in a wireless network merely links the abstract idea to a technological environment (MPEP 2106.05).
Step 2B: Claim 1 does not include any additional elements that amount to an integration of the judicial exceptions into a practical application, nor to significantly more than the judicial exceptions when taken alone or in combination. The additional elements describe receiving and transmitting data, which is considered well-understood, routine, conventional activity (MPEP 2106.05(d)(II)(I)) and generally link the abstract idea to a technological environment. Claim 1 is not patent eligible.
Regarding claim 2, the rejection of claim 1 is incorporated, further the claim recites: wherein the predefined prediction is performed by at least one of an expected inference time and a prediction accuracy, wherein the predefined prediction is set by at least one of a user, an electronic device, and a network device. This limitation amounts to more specifics of the mental process of determining whether the model meets a predefined prediction. Claim 2 is not patent eligible.
Regarding claim 3, the rejection of claim is incorporated, further the claim recites: receiving a feedback about performance of the at least one deployed candidate data driven model from the at least one target deployment environment, wherein the feedback comprises at least one of a detail about the performance, a number of cores, a central processing unit (CPU) frequency, a type of the CPU, a memory, an accuracy and an inference time. This limitation amounts to an additional element; more specifically, it’s considered insignificant extra-solution activity: data gathering. Claim 3 does not include any additional elements that amount to an integration of the judicial exceptions into a practical application, nor to significantly more than the judicial exceptions when taken alone or in combination. The additional elements describe receiving data, which is considered well-understood, routine, conventional activity (MPEP 2106.05(d)(II)(I)). Claim 3 is not patent eligible.
Regarding claim 4, the rejection of claim 1 is incorporated, further the claim recites: continuously learning a data item from the at least one deployed candidate data driven model from the at least one target deployment environment. This limitation amounts to linking the judicial exception to a technological environment: online/continual learning (MPEP 2106.05(g)). Additionally, the claim recites: and re-deploying an updated version of the at least one deployed candidate data driven model based upon continuously learning the data item. This limitation is considered insignificant extra-solution activity: data gathering. The claim does not include any additional elements that amount to an integration of the judicial exceptions into a practical application, nor to significantly more than the judicial exceptions when taken alone or in combination. The additional elements describe receiving data which is considered well-understood, routine, conventional activity and generally linking the judicial exception to online/continual learning. Claim 4 is not patent eligible.
Regarding claim 5, the rejection of claim 1 is incorporated, further the claim recites: wherein determining whether the at least one candidate data driven model meets the predefined prediction comprises: analyzing and predicting a performance the at least one candidate data driven model with respect to an inference time and a prediction accuracy for a continuous changing target deployment environment over a period of time; and determining whether the at least one candidate data driven model meets the predefined prediction based on the analyzes and prediction. This limitation amounts to more specifics of the mental process of determining whether a model meets a predefined prediction by analyzing and predicting the performance of the model which can still be done in the human mind. Claim 5 is not patent eligible,
Regarding claim 6, the rejection of claim 1 is incorporated, further the claim recites: wherein obtaining the at least one candidate data driven model from the plurality of candidate data driven models comprises: receiving at least one resource information associated with at least one data driven model and at least one target deployment environment to train and develop at least one candidate data driven model from the plurality of candidate data driven models; and obtaining the at least one candidate data driven model from the plurality of candidate data driven models based on the obtained information. This limitation merely describes the additional element of obtaining the model, which remains insignificant extra-solution activity: data gathering. The claim does not include any additional elements that amount to an integration of the judicial exceptions into a practical application, nor to significantly more than the judicial exceptions when taken alone or in combination. The additional elements describe receiving data which is considered well-understood, routine, conventional activity. Claim 6 is not patent eligible.
Regarding claim 7, the rejection of claim 6 is incorporated, further the claim recites: wherein the at least one resource information comprises at least one of a target network node identifier (ID), a number of core, a central processing unit (CPU) frequency, a type of the CPU, and a memory. This limitation merely describes the additional element of obtaining the model, which remains insignificant extra-solution activity: data gathering. The claim does not include any additional elements that amount to an integration of the judicial exceptions into a practical application, nor to significantly more than the judicial exceptions when taken alone or in combination. The additional elements describe receiving data which is considered well-understood, routine, conventional activity. Claim 7 is not patent eligible.
Regarding claim 8, the rejection of claim 1 is incorporated, further the claim recites: wherein tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven model comprises: selecting at least one data driven technique from a plurality of data driven techniques to predict an output within an expected inference time and a prediction accuracy; selecting at least one hyper-parameter from a plurality of hyper-parameters to predict the output within the expected inference time and the prediction accuracy; prioritizing the at least one data driven technique from the plurality of data driven techniques and the at least one hyper-parameter from the plurality of hyper-parameters; and tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven model based on the priority. This limitation amounts to more specifics of the mental process of tuning a hyper-parameter and data-driven technique which involves selecting a hyper-parameter and data-driven technique to tune with. Claim 8 is not patent eligible.
Regarding claim 9, the rejection of claim 1 is incorporated, further the claim recites: wherein tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven model is performed without deploying the at least one candidate data driven model. This limitation amounts to more specifics of the mental process of tuning a hyper-parameter and data-driven technique which is done before deployment. Claim 9 is not patent eligible.
Regarding claim 10, the rejection of claim 1 is incorporated, further the claim recites: wherein tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven model is performed to match an inference time and a prediction accuracy expected at the at least one target deployment environment. This limitation amounts to more specifics of the mental process of tuning a hyper-parameter and data-driven technique which is done to match an expected inference time and prediction accuracy. Claim 10 is not patent eligible.
Regarding claim 11, the rejection of claim 1 is incorporated, further the claim recites: wherein the at least one target deployment environment comprises at least one of an open centralized unit (O-CU), an open distributed unit (O-DU), a near-real-time radio access network (RAN) intelligent controller (RT RIC) (104), a non-RT RIC, a server, a cloud, an eNB, and a gNB. This limitation amounts to generally linking the judicial exception to a technological environment. The claim does not include any additional elements that amount to an integration of the judicial exceptions into a practical application, nor to significantly more than the judicial exceptions when taken alone or in combination. The additional element describes linking the judicial exception to a technological environment (MPEP 2106.05(g)). Claim 11 is not patent eligible.
Regarding claims 12-19, the claims recite essentially the same limitation as claims 1-10 with the addition of an electric device with a processor and memory coupled to a processor which falls into the statutory category of machine. The rejections of claims 1-10 are incorporated; claims 12-19 are not patent eligible.
Regarding claim 20, the claim recites essentially the same limitation as claim 1 with the addition of a non-transitory computer readable storage media which falls into the statutory category of manufacture. The rejection of claim 1 is incorporated; claim 20 is not patent eligible.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claim(s) 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over O-RAN.WG2.AIML-v01.03 (referred to as O-RANWG2) in view of Thornton (Auto-WEKA: Combined Selection and Hyperparameter Optimization of Classification Algorithms).
Regarding claim 1, O-RANWG2 teaches a method for training of a data driven model by an electronic device in a wireless network, the method comprising: obtaining at least one candidate data driven model from a plurality of candidate data driven models (Pg. 24 section 2, The ML designer or the SMO/Non-RT RIC will select and onboard the ML model and relevant meta data into the ML training host. Utilizing on the ML training data collection, the ML training host will initiate the model training. Once the model is trained and validated, it is published back in the SMO/Non-RT RIC catalogue); determining whether the at least one candidate data driven model meets a predefined prediction (Pg. 24 section 3, The ML designer can check whether a trained ML model from the SMO/Non-RT RIC catalogue can be deployed in the ML inference host for the given ML-assisted solution, by mapping the ML model and ML assisted solution requirements to performance capabilities discovered from Step 1, predefined prediction isn’t narrowly defined in the claim so the output of the model/ML assisted solution or the requirements can be the predefined prediction); deploying the at least one candidate data driven model to at least one target deployment environment upon determining that the at least one candidate data driven model meets the predefined prediction (Pg. 24 section 3, The ML designer can check whether a trained ML model from the SMO/Non-RT RIC catalogue can be deployed in the ML inference host for the given ML-assisted solution, by mapping the ML model and ML assisted solution requirements to performance capabilities discovered from Step 1. Upon successful validation, ML designer will inform the SMO/Non-RT RIC to initiate model deployment); and tuning(Pg. 25 section 6, Based on the feedback and data received from various MFs and actors, the SMO/Non-RT RIC can inform the ML designer that an update is required to the current model.)
O-RANWG2 fails to teach tuning at least one hyper-parameter and at least one data driven technique even though it specifies updating the model after receiving feedback, it does not explicitly discuss tuning hyper-parameters or the data driven technique.
Thornton teaches tuning at least one hyper-parameter and at least one data driven technique (Algorithm 1, Section 1, In what follows, we demonstrate that CASH can be viewed as a single hierarchical hyperparameter optimization problem, in which even the choice of algorithm itself is considered a hyperparameter. We also show that—based on this problem formulation—recent Bayesian optimization methods can obtain high quality results in reasonable time, algorithm 1 shows an iterative process to find an optimal set of hyper-parameters that includes algorithm selection, the data driven technique here is the learning algorithm).
Thornton is analogous to the claimed invention because it is in the field of optimizing model hyperparameters and data-driven techniques. O-RANWG2 is analogous to the claimed invention because it’s in the field of model deployment and updating in a wireless network. Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have used the optimization technique in Thornton along with the system in O-RANWG2 because O-RANWG mentions model optimization may involve optimizing hyperparameters and network structure while Thornton provides a technique to “yield better-performing models than standard algorithm selection and hyperparameter optimization methods” (Section 1).
Regarding claim 2, O-RANWG2 teaches wherein the predefined prediction is performed by at least one of an expected inference time and a prediction accuracy, wherein the predefined prediction is set by at least one of a user, an electronic device, and a network device (Pg. 23 section 1, a) Processing capability of the inference host (for example: dedicated resources such as CPU/GPU, memory etc. for ML model inference). b) Requirements on the ML assisted solutions such as execution time cap, delay sensitivity, Pg. 24 section 3, The ML designer can check whether a trained ML model from the SMO/Non-RT RIC catalogue can be deployed in the ML inference host for the given ML-assisted solution, by mapping the ML model and ML assisted solution requirements to performance capabilities discovered from Step 1).
Regarding claim 3, O-RANWG2 teaches receiving a feedback about performance of the at least one deployed candidate data driven model from the at least one target deployment environment, wherein the feedback comprises at least one of a detail about the performance, a number of cores, a central processing unit (CPU) frequency, a type of the CPU, a memory, an accuracy and an inference time (Pg. 24 section 5, The ML inference host and actors are expected to feedback or report the performance of the ML model to the SMO/Non-RT RIC the SMO/Non-RT RIC so that the SMO/Non-RT RIC can monitor the performance (e.g., accuracy, running time, network KPIs) of the ML model and potentially update the model or reselect the model to be executed.)
Regarding claim 4, O-RANWG2 teaches continuously learning a data item from the at least one deployed candidate data driven model from the at least one target deployment environment (Pg. 36, In deployment scenario 1.4, the AI/ML model is first offline trained by training host in the SMO/Non-RT RIC. Training host in the Near-RT RIC is used for continuous online learning); and re-deploying an updated version of the at least one deployed candidate data driven model based upon continuously learning the data item (Pg. 25 section 6, Based on the feedback and data received from various MFs and actors, the SMO/Non-RT RIC can inform the ML designer that an update is required to the current model. The ML designer will initiate the model selection and training step, but with the existing trained model. Once the model has been retrained, it will be deployed as described in Step 3, and the updated model will be used for ML inference).
Regarding claim 5, O-RANWG2 teaches wherein determining whether the at least one candidate data driven model meets the predefined prediction comprises: analyzing and predicting a performance the at least one candidate data driven model with respect to an inference time and a prediction accuracy (Pg. 23 section 1, a) Processing capability of the inference host (for example: dedicated resources such as CPU/GPU, memory etc. for ML model inference). b) Requirements on the ML assisted solutions such as execution time cap, delay sensitivity, Pg. 24 section 3, The ML designer can check whether a trained ML model from the SMO/Non-RT RIC catalogue can be deployed in the ML inference host for the given ML-assisted solution, by mapping the ML model and ML assisted solution requirements to performance capabilities discovered from Step 1) for a continuous changing target deployment environment over a period of time (Pg. 36, In deployment scenario 1.4, the AI/ML model is first offline trained by training host in the SMO/Non-RT RIC. Training host in the Near-RT RIC is used for continuous online learning); and determining whether the at least one candidate data driven model meets the predefined prediction based on the analyzes and prediction (Pg. 24 section 5, The ML inference host and actors are expected to feedback or report the performance of the ML model to the SMO/Non-RT RIC the SMO/Non-RT RIC so that the SMO/Non-RT RIC can monitor the performance (e.g., accuracy, running time, network KPIs) of the ML model and potentially update the model or reselect the model to be executed, continuous online learning where feedback is used to redeploy and the process of deployment involves mapping the ML model and ML assisted solution requirements to performance capabilities discovered which is essentially predicting the performance of the model, this mapping is considered predicting; additionally, the written description nor claim describe a more detailed form of prediction (Par 0102, At operation 910, the analytics engine (406) retrieves the allocated resource for the environment. At operation 912, the ML target validator (408) predicts if the tuned model will perform as per user requirement)).
Regarding claim 6, O-RANWG2 teaches wherein obtaining the at least one candidate data driven model from the plurality of candidate data driven models comprises: receiving at least one resource information associated with at least one data driven model and at least one target deployment environment to train and develop at least one candidate data driven model from the plurality of candidate data driven models (Pg. 23 section 1, a) Processing capability of the inference host (for example: dedicated resources such as CPU/GPU, memory etc. for ML model inference). b) Requirements on the ML assisted solutions such as execution time cap, delay sensitivity, Pg. 24 section 5, The ML inference host and actors are expected to feedback or report the performance of the ML model to the SMO/Non-RT RIC the SMO/Non-RT RIC so that the SMO/Non-RT RIC can monitor the performance (e.g., accuracy, running time, network KPIs) of the ML model); and obtaining the at least one candidate data driven model from the plurality of candidate data driven models based on the obtained information (Pg. 24 section 3, The ML designer can check whether a trained ML model from the SMO/Non-RT RIC catalogue can be deployed in the ML inference host for the given ML-assisted solution, by mapping the ML model and ML assisted solution requirements to performance capabilities discovered from Step 1).
Regarding claim 7, O-RANWG2 teaches wherein the at least one resource information comprises at least one of a target network node identifier (ID), a number of core, a central processing unit (CPU) frequency, a type of the CPU, and a memory (Pg. 23 section 1, a) Processing capability of the inference host (for example: dedicated resources such as CPU/GPU, memory etc. for ML model inference).)
Regarding claim 8, O-RANWG2 teaches to predict an output within an expected inference time and a prediction accuracy (Pg. 23 section 1, b) Requirements on the ML assisted solutions such as execution time cap, delay sensitivity).
Thornton teaches, which O-RANWG2 fails to teach, wherein tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven model comprises: selecting at least one data driven technique from a plurality of data driven techniques t(Algorithm 1, Section 1, We dub this the combined algorithm selection and hyperparameter optimization (CASH) problem, the algorithm selects a candidate configuration which contains the algorithm); prioritizing the at least one data driven technique from the plurality of data driven techniques and the at least one hyper-parameter from the plurality of hyper-parameters (Algorithm 1, the candidate configuration which is used to update the model which is interpreted as prioritization); and tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven model based on the priority (Section 3, 1) first builds a model ML that captures the dependence of loss function L on hyperparameter settings λ (line1 in Algorithm1). It then iterates the following steps: use ML to determine a promising candidate configuration of hyperparameters λ to evaluate next (line3); evaluate the loss c of λ (line4);and update the model ML with the new datapoint (λ,c) thus obtained (lines5–6)…SMBO then simply maximizes this function over Λ to select the most useful configuration λ to evaluate next, Algorithm 1, this process is tuning). The rationale for using the method of Thornton is explained in claim 1.
Regarding claim 9, O-RANWG2 teaches is performed without deploying the at least one candidate data driven model (Pg. 36, In deployment scenario 1.4, the AI/ML model is first offline trained by training host in the SMO/Non-RT RIC, Pg. 24 section 5, the SMO/Non-RT RIC can monitor the performance (e.g., accuracy, running time, network KPIs) of the ML model and potentially update the model or reselect the model to be executed, if training and updating can be done offline, so can the method described in Thornton). Thornton teaches, which O-RANWG2 fails to teach, tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven mode (Algorithm 1).
Regarding claim 10, O-RANWG2 teaches to match an inference time and a prediction accuracy expected at the at least one target deployment environment (Pg. 23 section 1, a) Processing capability of the inference host (for example: dedicated resources such as CPU/GPU, memory etc. for ML model inference). b) Requirements on the ML assisted solutions such as execution time cap, delay sensitivity, Pg. 24 section 5, The ML inference host and actors are expected to feedback or report the performance of the ML model to the SMO/Non-RT RIC the SMO/Non-RT RIC so that the SMO/Non-RT RIC can monitor the performance (e.g., accuracy, running time, network KPIs) of the ML model). Thornton teaches, which O-RANWG2 fails to teach, tuning the at least one hyper-parameter and the at least one data driven technique running in the at least one data driven mode (Algorithm 1).
Regarding claim 11, O-RANWG2 teaches wherein the at least one target deployment environment comprises at least one of an open centralized unit (O-CU), an open distributed unit (O-DU), a near-real-time radio access network (RAN) intelligent controller (RT RIC) (104), a non-RT RIC, a server, a cloud, an eNB, and a gNB (Pg. 26, For deployment scenario 1.1, where Non-RT RIC acts as inference host, termination/backup activation commands and acknowledgements would be communicated via SMO internal interface. On the other hand, for deployment scenario 1.2, in which AI/ML model is deployed in the Near-RT RIC).
Regarding claims 12-20, the claims teach essentially the same limitations as claims 1-11 or some combination thereof. Therefore, claims 12-20 are taught by O-RANWG2 in view of Thornton.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to NATNAEL A ASEGDEW whose telephone number is (571)270-0407. The examiner can normally be reached 7:30-5.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Kakali Chaki can be reached at (571) 272-3719. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/NATNAEL A ASEGDEW/ Examiner, Art Unit 2122
/MICHAEL H HOANG/ PRIMARY EXAMINER, Art Unit 2122
Prosecution Insights